Action classification apparatus, action classification method, and non-transitory storage medium
AssigneeNEC Corporation
InventorNoboru YOSHIDA
Life of this patent
Estimating the term…
Abstract
The present invention provides an action classification apparatus (10) including: an extraction unit (11) that extracts, from a moving image, a plurality of human movements presented in any number of frames; a time-series feature value computation unit (12) that computes a time-series feature value for the any number of frames by computing, for each of the extracted human movements, a feature value of a human pose in each of the any number of frames; a similarity computation unit (13) that computes a similarity between a plurality of the time-series feature values; and a classification unit (14) that classifies the plurality of extracted human movements, based on the similarity.
Claims20
An action classification apparatus comprising:
- at least one memory configured to store one or more instructions;
- and
- at least one processor configured to execute the one or more instructions to:
- extract, from a moving image, a plurality of human movements presented in any number of frames;
- compute a time-series feature value for any number of frames by computing, for each of the extracted human movements, a feature value of a human pose in each of the any number of frames;
- computes compute a similarity between a plurality of the time-series feature values;
- and
- classify a plurality of extracted human movements, based on the similarity.
An action classification method comprising,
- by a computer, executing:
- extracting, from a moving image, a plurality of human movements presented in any number of frames;
- computing a time-series feature value for any number of frames by computing, for each of the extracted human movements, a feature value of a human pose in each of the any number of frames;
- computing a similarity between a plurality of the time-series feature values;
- and
- classifying a plurality of extracted human movements, based on the similarity.
A non-transitory storage medium storing a program causing a computer to
- extract, from a moving image, a plurality of human movements presented in any number of frames;
- compute a time-series feature value for any number of frames by computing, for each of the extracted human movements, a feature value of a human pose in each of the any number of frames;
- compute a similarity between a plurality of the time-series feature values;
- and
- classify a plurality of extracted human movements, based on the similarity.
Dependencies are inferred from the claim text.